The present disclosure relates to techniques for calculating the geometric scale of an object. More specifically, the present disclosure relates to a technique for calculating the geometric scale or size of an object using two or more images of the object, which have different perspectives, and associated imaging-device characteristics (such as the focal lengths) when the images were acquired.
Thanks in part to an ever-expanding set of capabilities, electronic devices are becoming increasingly more versatile. This expanded functionality is increasing the popularity of a wide variety of applications, such as: virtual reality applications that provide artificial realities (which are sometimes referred to as ‘virtual worlds’), which project users into computer-generated spaces (such as a three-dimensional space); and augmented reality applications that provide live or indirect views of a physical environment (which is sometimes referred to as ‘augmented reality’) whose elements are augmented by superimposed computer-generated information (such as supplemental information and image(s) associated with a virtual reality application).
However, it can be difficult to incorporate content from images of the real world (such as an object in a digital image) into a virtual world or an augmented reality. In particular, in the absence of a frame of reference, it can be difficult to determine the size of objects in the images, which can complicate the scaling of the content in the images. As a consequence, many applications request that users manually increase or decrease the size of objects extracted from images before they are included in a virtual or augmented reality (and, more generally, in another image). For example, a user may be asked to scale an image of a couch so that it can be overlaid or included in a virtual environment of a house. (Alternatively, the user may be required to take a picture using predefined conditions, such as a predefined angle or distance from an object, which can be very impractical.)
This problem can degrade the usefulness of virtual reality and augmented reality applications, and other image-processing applications. As a consequence, the overall user experience is often degraded, which can frustrate application users.